X2 3800+ voltage and temps

[SkeeterUK]

Got a x2 3800+ skt 939 from ocuk about 2 months ago i guess. Overclocked it to 2.6ghz with a volts of 1.475 in bios (1.45 in everest reported). It was just getting too hot and so i turned the volts down a few notches, i did 1.450v and was 400% memtest in windows stable, and 3 hrs orthos small fft stable. I then lowered it again to 1.425 and again stable using tests for same time periods. Tho the darndest thing is the temps have remained or infact got a tiny bit worse. Used to idle while on desktop about 40 and sometimes now its like that still but can be 44 per core. And it was up to 50 per core once the other day and all i was doing was streamin vid stuff off net, wasnt gaming or anything. Tho last day or two has been very hot days for me so maybe thats why. At the mo i have a big 30cm fan blowing about the room, and a window open and temps are about 37/38c.

Btw when i take my side of case off (antec p150w) the temps go low as 33 per core. Does this mean my case air flow is not that good and should perhaps replace either front akasa ambere 120mm fan or the back of case fan which is a Aerocool turbine fan (led blue) 120mm which is nice and quiet tho doesnt push much air out back i think as its rated 37 CFM and rotates a quiet 950RPM

 

[SkeeterUK]

Nvm it froze with 1.425, it high pitched whined then turned itself off with 1.450 so back to 1.475

 

[Big.Wayne]

Hi SkeeterPSA,

Here are some results from a X2 3800+ I have been testing (LCB9E 0632DPMW)

Toledo™3800+ X2 @2600MHz - 1.30vCore - XP120 - Sharkoon 120mm '2000' - CPU 27°C-44°C Idle/Load

Toledo™3800+ X2 @2700MHz - 1.375vCore - XP120 - Sharkoon 120mm '2000' - CPU 27°C-46°C Idle/Load

I'm aware that some chips are better than others but I think 1.475v is too high volts for running 2600MHz? as you can see the chip I have will do this at 1.3vCore.

Also I'm not sure what cooler you are using but your CPU temps have scope to be improved.

Having a nice cool case always helps, if you take your case side off and the case temps drop quite a bit thats usually a good sign it needs some tweaking. Gotta get that warm/hot air away otherwise even the premium coolers will struggle to cool the chip.

Nvm it froze with 1.425, it high pitched whined then turned itself off with 1.450 so back to 1.475

I wonder what that means? Could that be the sound of some mosfets or caps squealing?

 

[SkeeterUK]

Asus bios set volts are usually not accurate so setting 1.475 like i have done its really 1.42-1.45v everest says. They usually undervolt u see.

I think my stepping is the same as yours but search is disabled so i cant confirm it.

The high pitch whine comes from speakers it happens when the comps froze or froze and about to turn off. Worse kind of instability.

Im on a freezer 64 pro, when touched when idle 40/40c temps, its warm but not hot that i cant touch. I applied a thin coating of mx-1 paste all over surface of cpu then a tiny bit and rubbed it on the heatsink then a tiny blob in center of cpu then put it in comp and fitted it all. Duno why its so warm tho but my room is boiling even with open window etc. North east of uk btw.

So duno what to do cos temp wise, its defo not in danger, max it gets to is 63 on one core when priming at 1.475v in bios (1.45v real) and the cpu is designed to be fine up to 65c. Normal operation is 38-41c ish so thats ok. Its just im thinking these temps maybe causing my pc to really heat my room up so i was hoping lowering volts would have helped but it didnt plus it proved instable. Either cpu doesnt like low volts when oc or maybe my motherboard i duno, its a asus a8v deluxe rev 2.0 1014 bios.


--

Edit heres my cpu, seems same as yours.

LCB9E 0632

ADA3800DAA5CD

Btw whats mosfets?

 

[Justintime]

I'm aware that some chips are better than others but I think 1.475v is too high volts for running 2600MHz? as you can see the chip I have will do this at 1.3vCore.

Had much much worse chips than that, first X2s i played with did'nt even do 2.5Ghz on less than 1.5v

 

[Big.Wayne]

a8v deluxe rev 2.0 1014 bios.

Hmm I don't know much about this chipset, not had a lot of luck with VIA motherboards. That board may well be holding back your chip, always used nForce 4 myself for Socket 939.

Btw whats mosfets?

This area on your motherboard. . .

This area tends to get hot when running an overclocked CPU, if its a multi-core the PWM gets 'a lot' hotter. On many boards there is a little heatsink in this area which helps disperse the heat, people then set-up a cooler to help this PWM heatsink work even better.

All the voltage to the chip passed through this PWM area and when it starts to get too hot the voltage to your CPu can fluctuate too much. Having said that the PWM can safely run at a very high temp but I have always thought its better to keep it cool as possible.

Any of the suggestions below could be the reason your chip is being held back?

Bad CPU cooling?
Bad Mosfet cooling?
Bad motherboard?
Bad chip?

Had much much worse chips than that, first X2s i played with did'nt even do 2.5Ghz on less than 1.5v

Fair point, although you would think a recently purchased chip would do better?

 

[geeza]

mines at 2.8

46 and 41c idle. Full load goes to 90 sometimes but as long as my computer doesnt set of fire or not turn on again im not too bothered

cpu-z says 1.312, but cpu-z is lieing i think

 

[GeForce]

Full load goes to 90 sometimes

Toasty.

 

[Big.Wayne]

mines at 2.8

Full load goes to 90 sometimes

im not too bothered

Post some stability and volts/temps monitoring screenies pls.

Also what cooling are you using?

 

[SkeeterUK]

You sure you dont mean 60 on load not 90?

 

[Justintime]

Fair point, although you would think a recently purchased chip would do better?

Depends, how old stock was and then its just down to plain luck. Take for example during the AXIA craze i had a 1Ghz that would'nt do 1.2Ghz without 1.85V or more, sold that in a client rig at stock and the next one did 1.4Ghz on 1.75v (stock), eventually got one that i kept for years as she did 1.8Ghz on 1.85v on water and later on on air with an Alpha Pal 8045.

 

[Angilion]

mines at 2.8

46 and 41c idle. Full load goes to 90 sometimes but as long as my computer doesnt set of fire or not turn on again im not too bothered

cpu-z says 1.312, but cpu-z is lieing i think

44-49C delta from idle to full load is far too high. Your cooling isn't effective enough.

Says a lot for how heat-tolerant the chip is, though. The suggested max is in the high 60s, as far as I know.

 

[ghgh]

x2 3800+ @ 2700MHz , 1.34v

full load temperature: core 0 - 54c : core 1 - 58c.

not too bothered about the temps as it is stable as it can be.

ive had it stable at 3.06GHz with 1.55v but never had RAM to justify it.

 

[Amonlym]

could well be the chipset going funny, my AX8 (kt890) does wierd things on high htt.

try setting the ram on a 166mhz divider, if you haven't already, if you have try setting it a little lower still.

try upping the nb voltage a touch

try moving any pci cards you have to different slots

try dropping 1/2mhz fsb and see if that fixes it. - i can run at 273*11 for 3ghz @ 1.55V but it isnt exacty stable. 250*12 works perfectly. dropping to 243htt lets vcore drop to 1.4v from 1.55!

 

[V|per]

x2 3800+ @ 2700MHz , 1.34v

full load temperature: core 0 - 54c : core 1 - 58c.

not too bothered about the temps as it is stable as it can be.

ive had it stable at 3.06GHz with 1.55v but never had RAM to justify it.

Run the mem on a devider

The AMD's have an onboard memory controller, and it is generally believed that the system with benefit from tighter timings than sheer mhz speed of the ram...

 

[ghgh]

Run the mem on a devider

The AMD's have an onboard memory controller, and it is generally believed that the system with benefit from tighter timings than sheer mhz speed of the ram...

I did, but the 133 divider doesnt work on this motherboard so I had to put it on the 100MHz divider - what I found was that SuperPI was slower at 3GHz with the same on the 100MHz divider than when the CPU is at 2.7GHz on the 166MHz divider.

 

[SkeeterUK]

Divider might work if you put a bit more volts on the ram.

 

[V|per]

I did, but the 133 divider doesnt work on this motherboard so I had to put it on the 100MHz divider - what I found was that SuperPI was slower at 3GHz with the same on the 100MHz divider than when the CPU is at 2.7GHz on the 166MHz divider.

Thats because superPI is one of the only programs that benefits from sheer memory speed.

If the cpu's FSB is at 300mhz, you'll want the 133 edvider to get the memory at 200mhz (i assume you are using pc3200) if it's being a bit daft, just nudge the voltage to the ram a little, try loosening the timings slightly to see if it helps too.

 

[ghgh]

Thats because superPI is one of the only programs that benefits from sheer memory speed.

If the cpu's FSB is at 300mhz, you'll want the 133 edvider to get the memory at 200mhz (i assume you are using pc3200) if it's being a bit daft, just nudge the voltage to the ram a little, try loosening the timings slightly to see if it helps too.

thanks for the help.

the RAM doesnt POST at CAS 3.0 so I cant losen the timings and my board only lets me set 2.8v on the RAM maximum so I cant up that and would you belive it, the 133MHz RAM divider just simply doesn't POST or it hangs on the BIOS screen forever. I tried everything trying to get 3006MHz faster than my 2700MHz but it won't work. Yeah, im using Geil Value PC3200 (only 2 x 256MB) on the x2. Will try it with my BH5 when I get a DFI.

cheers,
ghgh